arrange_at doesn't work when select Chinese colnames via select helpers

[shizidushu]

It seems arranged_at doesn't work properly if Chinese colnames used in select helpers.

library(dplyr)
#> 
#> Attaching package: 'dplyr'
#> The following objects are masked from 'package:stats':
#> 
#>     filter, lag
#> The following objects are masked from 'package:base':
#> 
#>     intersect, setdiff, setequal, union
  df <- data.frame(qty = c(3,1,5,4), 金额 = c(1,3,4,2))
  arrange_at(df, vars(ends_with('金额')))
#>   qty 金额
#> 1   3    1
#> 2   4    2
#> 3   1    3
#> 4   5    4
  arrange_at(df, desc(vars(ends_with('金额'))))
#>   qty 金额
#> 1   3    1
#> 2   4    2
#> 3   1    3
#> 4   5    4
  sessionInfo()
#> R version 3.5.1 (2018-07-02)
#> Platform: x86_64-pc-linux-gnu (64-bit)
#> Running under: Debian GNU/Linux 9 (stretch)
#> 
#> Matrix products: default
#> BLAS: /usr/lib/openblas-base/libblas.so.3
#> LAPACK: /usr/lib/libopenblasp-r0.2.19.so
#> 
#> locale:
#>  [1] LC_CTYPE=en_US.UTF-8       LC_NUMERIC=C              
#>  [3] LC_TIME=en_US.UTF-8        LC_COLLATE=en_US.UTF-8    
#>  [5] LC_MONETARY=en_US.UTF-8    LC_MESSAGES=C             
#>  [7] LC_PAPER=en_US.UTF-8       LC_NAME=C                 
#>  [9] LC_ADDRESS=C               LC_TELEPHONE=C            
#> [11] LC_MEASUREMENT=en_US.UTF-8 LC_IDENTIFICATION=C       
#> 
#> attached base packages:
#> [1] stats     graphics  grDevices utils     datasets  methods   base     
#> 
#> other attached packages:
#> [1] bindrcpp_0.2.2 dplyr_0.7.6   
#> 
#> loaded via a namespace (and not attached):
#>  [1] Rcpp_0.12.18     crayon_1.3.4     assertthat_0.2.0 digest_0.6.16   
#>  [5] rprojroot_1.3-2  R6_2.2.2         backports_1.1.2  magrittr_1.5    
#>  [9] evaluate_0.11    pillar_1.3.0     rlang_0.2.2.9000 stringi_1.2.4   
#> [13] rmarkdown_1.10   tools_3.5.1      stringr_1.3.1    glue_1.3.0      
#> [17] purrr_0.2.5      xfun_0.3         yaml_2.2.0       compiler_3.5.1  
#> [21] pkgconfig_2.0.2  htmltools_0.3.6  tidyselect_0.2.4 bindr_0.1.1     
#> [25] knitr_1.20.15    tibble_1.4.2

[batpigandme]

Hi @shizidushu,

I don't think this has to do with the character encoding, as the same thing occurs if you do the same thing using vars(ends_with('ty'))

library(tidyverse)
df <- data.frame(qty = c(3,1,5,4), 金额 = c(1,3,4,2))

# normal arrange descending
arrange(df, desc(`金额`))
#>   qty 金额
#> 1   5    4
#> 2   1    3
#> 3   4    2
#> 4   3    1

# works
arrange_at(df, vars(ends_with('ty')))
#>   qty 金额
#> 1   1    3
#> 2   3    1
#> 3   4    2
#> 4   5    4

# works
arrange_at(df, vars(ends_with('金额')))
#>   qty 金额
#> 1   3    1
#> 2   4    2
#> 3   1    3
#> 4   5    4

# works
arrange_at(df, vars(ends_with('ty')))
#>   qty 金额
#> 1   1    3
#> 2   3    1
#> 3   4    2
#> 4   5    4

# doesn't work
arrange_at(df, desc(vars(ends_with('ty'))))
#>   qty 金额
#> 1   3    1
#> 2   4    2
#> 3   1    3
#> 4   5    4

Created on 2018-09-12 by the reprex package (v0.2.0.9000).

[yutannihilation]

desc() cannot be applied to vars(), which specifies the column indices, not expressions. If you want to transform the variables, the transformation functions should be specified on .funs argument.

library(dplyr, warn.conflicts = FALSE)

df <- data.frame(qty = c(3,1,5,4), 金额 = c(1,3,4,2))

arrange_at(df, vars(ends_with('金额')))
#>   qty 金额
#> 1   3    1
#> 2   4    2
#> 3   1    3
#> 4   5    4

arrange_at(df, vars(ends_with('金额')), desc)
#>   qty 金额
#> 1   5    4
#> 2   1    3
#> 3   4    2
#> 4   3    1

Created on 2018-09-13 by the reprex package (v0.2.0.9000).

FYI, Examples section of arrange_all explains this:

# You can supply a function that will be applied before taking the
# ordering of the variables. The variables of the sorted tibble
# keep their original values.
arrange_all(df, desc)
arrange_all(df, funs(desc(.)))

[shizidushu]

@batpigandme @yutannihilation Thanks for your reply.

So .funs do the transformation and the transformed values are used to generate the order.

But now I'm confused about what .funs do when I specify two functions.

library(tidyverse)
df <- data.frame(qty1 = c(1, 1, 2), qty2 = c(-1, 0, -2), amt = c(10, 13, 11))
# arrange_at and their arrange equivalents?
arrange_at(df, vars(starts_with('qty')), funs(desc, identity)) # arrange(df, desc(qty1), qty2)
#>   qty1 qty2 amt
#> 1    2   -2  11
#> 2    1    0  13
#> 3    1   -1  10
arrange_at(df, vars(starts_with('qty')), funs(identity, identity)) # arrange(df, qty1, qty2)
#>   qty1 qty2 amt
#> 1    1   -1  10
#> 2    1    0  13
#> 3    2   -2  11
arrange_at(df, vars(starts_with('qty')), funs(identity, desc)) # arrange(df, qty1, desc(qty2))
#>   qty1 qty2 amt
#> 1    1   -1  10
#> 2    1    0  13
#> 3    2   -2  11

Created on 2018-09-13 by the reprex package (v0.2.0).

[yutannihilation]

IIUC, arrange_at(df, vars(starts_with('qty')), funs(desc, identity)) is equivalent to

arrange(df, desc(qty1), decs(qty2), identity(qty1), identify(qty2))

If you want to transform variables separately, I recommend you to use tidyeval...

[shizidushu]

@yutannihilation Thanks. I got it.

[lock[bot]]

This old issue has been automatically locked. If you believe you have found a related problem, please file a new issue (with reprex) and link to this issue. https://reprex.tidyverse.org/